About a month ago, the symposium on my book over at The Disorder of Things that I mentioned in an earlier post reached a critical stage: posts by Paul, Joe, Nick, and Meera were up, and it was now my turn. And then life intervened, and lo and behold it’s been almost a month and I haven’t put up a reply. I finally got a chance to start work on a reply this weekend, but it’s grown so large in conception that I am going to have to split it into three parts. The first part is below the fold in this post; the second and third parts will follow shortly, depending — in all honesty — on how much time I get on the plane (and waiting in airports for connecting flights) to and from Montreal for the ISA this week. There are themes I introduce in this first part that I will take up again in subsequent parts, so bear that in mind.

Let me just say at the outset of my reply to the gang of four at the Disorder of Things that the way they have engaged with my book is not only the precise kind of reaction I hoped to provoke — the discussion is incredibly rich and subtle, and if C of I does nothing other than provide the occasion for discussions of this quality then I will consider the book a successful intervention — but it is also the sort of careful reading of one’s work that constitutes a rare gift. To be taken that seriously and read that closely is something that we scholars dream about, especially while in graduate school before the everyday actuality of building one’s career (hopefully while still living one’s life!) sets in and one’s colleagues never seem to have enough time to really sit down and engage that profoundly with one’s work — and to be fair and honest, you don’t have time to read their work that closely either. But this just makes me all the more grateful for the opportunity to engage in this discussion, as I now appreciate its value much more than I might have otherwise.

Not to sound like a grizzled old geezer, either, but I do want to make another preliminary remark that concerns ages and career stages. I am fairly certain that I would not have been capable of writing this book a decade ago, when I was fresh out of graduate school and still completely convinced — my substantive sympathies with Nietzsche and Weber and Wittgenstein to the contrary, but it took several years for that tension to actually work itself out in my scholarly practice — that a good argument would and could carry the day in a dispute among scholars. It took me a while to actually realize that while a good argument is necessary for anyone inhabiting the scholarly form of life (because a poor argument is a liability and a weakness in this game), it is far from sufficient. Even the best of arguments — and I want to be clear here, “best” in my lexicon doesn’t necessarily imply any kind of binding universal standards for what makes for a sound argument, but is fully compatible with the position I actually accept, which is that the conditions of argumentative soundness are both locally situated and practically embedded — can fail to carry the day in the face of institutionalized practices and organizations that instantiate quite different principles. Even in academia, which was a realization that took a while to dawn on me; only a combination of career experience and a conceptual vocabulary that seemed to order those experiences for me in a useful way (shout-out here to Andrew Abbott’s brilliant book Chaos of Disciplines, and to Weber’s classic vocation essays, which I’d read before but didn’t really appreciate until I had the requisite practical puzzles that they helped me resolve) led me to really comprehend that the practical value of a good argument was less about what it said and more about what role it played and what effects it had. Which, to reiterate, is not a warrant for bad arguments, but it is a warrant for arguments that self-consciously start not from some idealized realm of pure intellectualism but from the actual place — conceptually, institutionally, even culturally — where you and your audience are located.

Accordingly, C of I is not, and is not intended to be, a definitive account of what it means to study world politics scientifically. In fact, it’s not even an argument that one should study world politics scientifically, but is more of an intellectual exercise that takes the form: “supposing that one wanted to study world politics scientifically, how would one go about doing that?” Several of my interlocutors pick up on this point and I will have more to say about it in a moment, but I think it important to establish here at the beginning of my response that there are two independent but related lines of critique and evaluation of the book: the internal coherence of the argument itself, and the implications of the argument as an intervention into the currently-constituted scholarly field of global IR. These two dimensions are irreducible to one another, I believe, and the main difference between the me of a decade ago and the me who wrote C of I is that I think I have a clearer appreciation of the differences — and that I have written a book that privileges its status as an intervention over its status as a pure argument. There are areas I did not choose to explore and issues I did not choose to tackle, not because they weren’t logically entailed by the arguments in the book (and not because I don’t have opinions about them!), but because I’d rather spark a conversation than foreclose any discussion.

All of that said, I’d like to proceed by taking up three categories of issue raised by these four extremely rich and detailed engagements. First, at the risk of reiterating things I said in the book, I want to begin by revisiting two pieces of my conceptual apparatus that my interlocutors raise critical questions about: the distinction between scientific and philosophical ontology, and the category “science” itself. Second, I want to respond to the challenges to my methodological lexicon raised in the name of ethics and (transformative) politics. Third, I want to take up the issue of methodological distinctions, and simultaneously defend both my opposition to the eclectic mixing of methodologies and my advocacy of engaged pluralism, even though an engaged pluralism may well make the methodological distinctions I disclose completely irrelevant to future global IR scholarship. Indeed, I hope that they do become irrelevant eventually, because this would mean that they and the book as a whole have done their work in provoking a global IR field that has moved on to other ways of articulating those concerns that it takes in from the wider culture and subjects to the distinctively scholarly exercises of logical elaboration and knowledge-production.

I. standing where I stand

One of points that I articulate in C of I but do not spend much time defending is that philosophical ontology should be distinguished from scientific ontology: that the “hook-up” between the mind and the world is logically disjointed from any particular conceptualization of the objects in the world, and that methodological discussion pertains to the former rather than to the latter. My position is thus to be distinguished both from the neopositivist claim to be focusing on the world rather than being distracted by irrelevant philosophical concerns (a claim that only makes sense if we presume the methodological correctness of neopositivism, a presumption that would make methodological discussion superfluous), and from the critical realist claim that (scientific) ontology comes before methodology (a claim that, as Fred Chernoff among others has also argued, runs the very real risk of turning into a form of ontological dogmatism). I do not argue that philosophical ontology ought to set limits on scientific ontology, as any particular substantive conception of what exists can in principle be cashed out according to the strictures of any methodology, and can thus be incorporated into a wide variety of potential research projects. To claim otherwise would be to claim the ultimate irrelevance of purely methodological discussion and a corresponding need to focus attention on substantive claims — a position close to the Feyerabendian stance adopted by Paul and seemingly underlying the continuum of presumed ontological stability proposed by Meera.

It strikes me that this position makes two rather important presumptions: first, that the world has a definite and determinate character which we could get more or less right, and second, that arguments about the epistemic warrant for empirical claims have to be consistent with the substantive content of those claims. The first argument is nothing other than the presumption of mind-world dualism, which — as I try to illustrate at considerable length in the book — is certainly a contestable proposition, given the existence of more or less well-developed mind-world monist positions. Contra Joe’s claim, though, this does not make my overall argument a monistic one; indeed, if I maintained that mind-world monism were simply true then I would have written a very different book, one that did not make so much of the distinction between monism and dualism as a way of organizing IR methodological debates. It is enough that we be unable to resolve the issue in favor of either monism or dualism, at least at this point in time, for my overall argument to have some traction. If we could resolve the issue, either by philosophical fiat or by an implicit consensus about what it would mean for some claim about the world to be true (and parenthetically, Feyerabend’s free-wheeling methodological anarchism seems tacitly underpinned, as many if not most arguments in the philosophy of natural science seem to be, by a broad consensus that natural science works and is therefore in some sense true, and that it is this working that has to be explained), that would indeed leave us free to focus on substantive propositions, and to divide up our theories accordingly. But since we cannot (and I will return to Nick’s ingenious attempt to do so below, and in subsequent parts of my reply), and since different methodological perspectives harbor and afford different kinds of warrants for substantive propositions whatever their content, I think that we need to treat philosophical ontology separately and allow the question of whether the world has a definite and knowable character to remain endogenous to the ensuing discussion.

So much for the issue of reducing philosophical ontology to scientific ontology. But what of the reverse claim, that different philosophical ontologies covertly entail particular kinds of scientific ontologies? Underlying both Paul and Meera’s positions is, I believe, a value-commitment that I once thought a lot more highly of than I do now — the claim that methodological assumptions have to be consistent with substantive assumptions. I first encountered this point in Trevor Barnes’ excellent book Logics of Dislocation, where it appears as a criticism of the rational actor model and offers two choices: either people need oversimplifying assumptions like the rational actor model itself, in which case their heuristics and not the rational structure of their situations determines their behavior and rational choice theory fails, or people can react rationally to the actual situations in which they find themselves, in which case there would be no need for scholars to use admittedly-oversimplified rational actor notions to explain outcomes and again rational choice theory fails. And it can also be found in the opening of Robert Jervis’ Perception and Misperception in International Politics, in the form of a claim that scholars should assume that policymakers learn about the world in more or less the same way that scholars do. In both cases, the argument turns on the claim that how scholars generate knowledge and how people in their everydayness generate knowledge are fundamentally similar, and so accounts of scholarly methodology need to regard scholars as equally subject to the models and concepts with which they work — in Donald Moon’s dated and sexist but still useful language, that the “model of man” with which a theorist operates has to also apply to the theorist him- or herself.

But whatever its rhetorical effectiveness — which derives, I think, in no small part from the apparent humanism of equating scholars and their objects of study — this argument is not, I think, actually all that logically compelling. A neopositivist is not required to treat the subjects of her or his study as hypothesis-testers; indeed, a neopositivist ought to make that claim (“people in their everydayness are hypothesis-testers”) a hypothesis and test it, and if the hypothesis were to be falsified, that would have precisely no implications for whether the investigator should remain a neopositivist. An analyticist could simultaneously argue that knowledge is produced through the creative deployment of cultural resources and value-commitments and deploy a model of rational action that featured people doing no such thing; indeed, as I argue in the book, properly understood this is precisely what rational choice theory actually does or should do. By the same token, an analyticist could deploy a model of “situated creativity” that more closely mirrored the presumptions that the methodology makes about how scholarly knowledge is produced, but this would be strictly speaking an accidental correspondence — much like the fact that Roy Bhaskar uses his realist theory of science as the inspiration for his more general account of human society. But there is no reason that adopting either of those scientific ontologies would commit one to the philosophical ontology that they most resemble, just as there is no necessary reason why all critical realists would or should adopt Bhaskar’s particular account of the social world. Reflexive theory is the harder case, because it is perhaps easier to imagine the appeal to consistency between scholarly knower and everyday knower gaining traction in such a methodology, most theorists working in that tradition maintain that a self-located scientific knower has a better grasp of things than an ordinary everyday knower … so here again there is no need for consistency between philosophical and scientific ontologies. There may well be what Weber would call “elective affinities” between positions on both registers — most researchers who use participant-observation techniques to elucidate intersubjective cultural lifeworlds populated by meaning-making human beings are monists, and most researchers seeking to correct the operative misperceptions of their subjects by introducing novel information or presenting disconfirming evidence are dualists — but that’s not a strictly logical entailment. It’s almost certainly a sociological one.

Which brings me to the other point I want to make about the distinction between philosophical and scientific ontology, and about my retaining of the category “science” as the sign under which to conduct my analysis of how inquiry is and should be conducted in IR. In a field dominated as ours is by neopositivist ways of knowing, starting with substantive propositions about the world as a way of organizing theories and debates leaves unchallenged the notion that all such propositions need to be subjected to a neopositivist mode of evaluation, which means that they have to be converted into falsifiable propositions and tested against appropriate data. This appeals to neopositivists, obviously, but not necessarily to the rest of the field. The point can be generalized: only if there is methodological consensus then we we can safely focus on substance and ignore methodology, and indeed, this happens in focused research communities throughout the field. But one of the bizarre things about the way we organize IR discussions is that we usually just name the substantive focus and omit any mention of the fact that, for instance, the “democratic peace” research community is pretty much entirely populated by neopositivists, and that any consensus that the community reaches is therefore necessarily underpinned by that methodological consensus. And this in turn gives rise to the misleading position that studying a particular substantive area requires a certain kind of methodology, which is more likely than not a neopositivist methodology (except in those few research communities where neopositivism is explicitly rejected — but it’s no better to argue, or more typically, assume that the study of non-traditional security threats or religion or gender hierarchies or whatever demands a non-neopositivist methodology). Separating scientific and philosophical ontology — substantive claims and methodology — explodes these more or less purely sociological and philosophically accidental arrangements, and opens up new lines of research that combine methodology and substantive theory in novel ways, and for that reason alone I think the separation is compelling.

Something similar could be said of my very deliberate decision to retain the category “science” in the book, rather than switching to “inquiry” or just saying “valid knowledge.” While I am quite sympathetic to the argument that because the term comes with so much cultural and institutional baggage it may not be the best guide to thinking clearly about methodological matters, it is precisely that cultural and institutional baggage that (in my view) makes it unavoidable. Because the claim to science functions as such a trump-card in discussions of methodology — particularly in the study of social life, including world politics — ignoring it or rejecting it seems most unwise, approximately the rhetorical equivalent of placing oneself on the tracks directly in front of a speeding train and asking it politely to stop. “Science” is a term to conjure with in our field, and although philosophers of science like Larry Laudan can propose abandoning the term in order to simply focus on warranted knowledge-claims, I don’t think that we can actually do this in global IR because the claim to being a science is so much a part of the way that the field is structured: the hierarchies of journals, the prestige of doctoral programs, and the evaluation of who is and who is not suited for permanent positions in the academy (to name only three sociological elements) are all wrapped up with the science question. So a frontal assault makes more sense than simply abandoning the struggle, something Joe acknowledges in his concern about whether ethical inquiry belongs in the domain of legitimate knowledge: the notion of science is powerful, and is often used to shut down (“discipline”) empirical inquiry in all sorts of ways that discourage rather than encourage contentious conversations.

Aside from this sociological and tactical consideration — which, I hasten to add, should not be mistaken for anything like a positive argument in favor of being scientific in our approach to the study of world politics; my overall project in this book is, I think, better read as “if you want to be scientific, here is a discussion about what that might mean in the study of world politics,” and yes I am very aware of the rhetorical duplicity but also the practical utility of this move since “being scientific” is more than just one among other intellectual choices — there is a broader philosophical argument about the category and the label “science” that informs my overall project in the book. That, of course, is that science as a category is plural and pluralist, that “the scientific method” is a misnomer that is perhaps appropriate to elementary-school exercises in laboratory experimentation but nowhere else, and that there is no compelling philosophical reason for an IR scholar to reject the label “science” just because she or he does not engage in large-n quantitative hypothesis-testing. Nick suggests that my elaboration of four ways of being scientific threatens to undermine the whole “scientific project,” and notes that partisans of each of my four methodological options are likely to regard arguments unfolding in accord with any of the other three methodological stances as not really being scientific, but I don’t think that either of those points (even though they are good points) actually militate against retaining the label “science.” Yes, partisans of a particular methodology might claim their way of worlding as uniquely scientific, but they would be wrong inasmuch as partisans of other methodologies would claim exactly the same thing — and those claims cannot be adjudicated, but rely in practice on a leap of faith or constitutive commitment about the mind-world “hook-up.” Sure, if Robert Brandom or Bruno Latour were correct about the mind-world “hook-up,” that would matter a great deal, but a) I am unsure how you would ever actually come to that conclusion without tautologically presupposing a position on the mind-world “hook-up” in the first place, and b) although I am no expert on Brandom, he sounds like an analyticist to me, and Latour is certainly and unequivocally not a dualist (although one must be very careful here not to confuse Latour’s notion of the non-human and non-language-using actant with a philosophically-ontological dualism; if Latour is any kind of a dualist he’s a dualist at the level of scientific ontology, and I’d actually say that he’s a pluralist at that level in the William James sense of a pluralistic universe … but I digress). The problem, then, is that everyone claims to be right even though there is no non-tautological way to establish their own rightness, and the solution — at least the solution I propose — is that everybody take it down a notch, and recognize that other perspectives are as skeptical about your position as you are about theirs.

As for “the scientific project,” I’m not convinced that there actually is such a thing; there are scientistic ideologies, to be sure, but there is no more “the” science than there is “the” art or “the” music. We have no problem speaking of forms or modes of art and music, but we seem more uncomfortable with the idea that there might be forms of science even though, empirically speaking, we do have multiple forms of scientific endeavor in practice. I find this strange. From the various efforts to define science one finds in the philosophical literature, I extracted three of the most widely-used and uncontroversial characteristics — systematicity, publicity, and worldliness — as a way of re-visioning the notion of science in our field in a way that would hopefully prevent its misuse. In particular I hope that this definition allows us to speak of forms of the scientific study of world politics more easily, without falling into the misconception that all scientific studies have to be somehow compatible with one another and easily subsumable under a common set of claims. Contra Nick, I actually think it’s a strength of my conception of science that it contains and can support no single definition of “progress,” since that allows different methodologies to specify different criteria for evaluating arguments and improving them — and because I can’t imagine any non-tautological way of actually evaluating different conceptions of the mind-world “hook-up” and thus achieving “progress” on that score. One makes the assumptions of different methodologies compete with one another directly only by waging holy wars; that leaves only divorce, or a contentious but indirect dialogue between different methodological perspectives and the insights that they generate, as less fundamentalist options. I’m not a big fan of divorce because I’m well aware that partisans of any methodological position need the existence of other positions to remind them of the partial and incomplete nature of their own position, so contentious dialogue as a form of therapy for the error of mistaking one’s own views with The One True Way That The World Is strikes me as a much better alternative. Should there be philosophical discussion of methodology in IR? Yes, of course there should be, but we shouldn’t kid ourselves that we’re going to actually resolve any of these issues. Resolving philosophical controversies is not our job, especially if we regard our job to be that of a social scientist; particularly in that case, our job in that case is to generate valid knowledge about world politics, which is a big enough job on its own.

Of course, much of what I’ve said here depends on two propositions: 1) we cannot resolve controversies of value-commitments, including and perhaps especially value-commitments about philosophical ontology, scientifically; and 2) science broadly conceived has the potential to generate the kind of interesting empirical results I am calling for. In the next part of my reply I want to take up both of those points, by way of clarifying and defending my broad notion of science vis-a-vis both ethics and politics as separate fields of human endeavor.

Next: part the second, “ethics and politics, which are good things but not science,” to be followed by part the third, “don’t cross the streams, but do talk about doing so.”